Commit Graph
4041 Commits
Author SHA1 Message Date
wucm667 27de27ac64 fix(risk-control): remove deleted groups from moderation scope 2026-09-29 23:24:45 +08:00
Wesley Liddick fe9dadfc7d Merge pull request #7679 from PMExtra/feat/risk-control-allowlist
feat: 新增可信用户风控白名单,保留审计并免除本地自动处罚
2026-09-29 15:53:18 +08:00
Wesley Liddick 052e08e764 Merge pull request #7684 from william-drakemond/feat/claude-reset-credits-ui
feat(claude): 查询 Claude 重置次数(只读),UI 复用 Codex 次数按钮 / Read-only Claude reset credit count, reusing Codex count UI
2026-09-29 15:50:53 +08:00
Wesley Liddick 1f955af353 Merge pull request #7380 from wucm667/fix/issue-7377-forward-stream-usage
fix(gateway): forward received chat stream usage
2026-09-29 15:44:10 +08:00
Wesley Liddick ccb027b515 Merge pull request #7378 from wucm667/fix/issue-7368-composite-ws-routing
fix(gateway): resolve composite routes for websocket aliases
2026-09-29 15:43:52 +08:00
Wesley Liddick 883a53fea0 Merge pull request #7643 from wucm667/fix/issue-7637-antigravity-compat-keepalive
fix(antigravity): keep compat streams alive before first content
2026-09-29 15:43:19 +08:00
Wesley Liddick b35a611538 Merge pull request #7646 from wucm667/feat/issue-7642-dashboard-spending-trend
feat(dashboard): toggle recent usage between tokens and spending
2026-09-29 15:43:07 +08:00
Wesley Liddick 1dca86f287 Merge pull request #7701 from wucm667/fix/issue-7699-tool-rewrite
fix(anthropic): rewrite tool names in one body pass
2026-09-29 15:42:45 +08:00
wucm667 27a9421e6a fix(anthropic): rewrite tool names in one body pass 2026-09-29 15:00:16 +08:00
William DrakeandClaude Opus 5.5 860782e6fb fix(claude): harden reset credit projection and cell state (review R2)
- Drop selection_token (unsalted hash of a short grant ID, reversible) and
  claudeGrantSelection from the read-only DTO; assert it never serializes.
- query() returns only the projected result; no raw upstream block escapes.
- Clear a cooldown_until that is already past, using the same `now` as the
  redeemable check; the cell also shows the cooldown only when in the future.
- Invalidate in-flight queries when account id, platform or type changes.
- Sort the earliest-expiry fallback by parsed timestamp, invalid/missing last.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-09-28 20:59:39 -04:00
korkin25 48970f61b4 fix(claude): explicitly close status HTTP response 2026-09-28 18:04:14 -04:00
korkin25 6a8f921650 feat(claude): expose sanitized native reset credit status 2026-09-28 18:04:13 -04:00
haruka 2cc459e24f style: satisfy Sonnet 5.5 lint checks 2026-09-29 05:20:04 +08:00
haruka 8490a81867 feat: support Claude Sonnet 5.5 2026-09-29 04:55:07 +08:00
PMExtra 6a69dd051a feat: add risk control user allowlist 2026-09-29 01:50:21 +08:00
wucm667 aa3a3ea787 feat(dashboard): toggle recent usage between tokens and spending 2026-09-28 14:48:42 +08:00
wucm667 bd48b2b2d2 fix(antigravity): keep compat streams alive before first content 2026-09-28 12:15:27 +08:00
Wesley Liddick 4c00df2e01 Merge pull request #7638 from Achordchan/fix/anthropic-structured-output-beta
fix(anthropic): preserve requested structured outputs beta
2026-09-28 10:52:52 +08:00
Wesley Liddick b5298fdde8 Merge pull request #7538 from TouHouQing/fix/claude-thinking-disabled-openai
fix(openai): honor disabled thinking in Claude Messages bridge
2026-09-28 10:51:30 +08:00
Wesley Liddick 5a350108c4 Merge pull request #7607 from ResidualBlood/fix/antigravity-malformed-function-call-retry
fix(antigravity): trigger failover retry on empty stream caused by MALFORMED_FUNCTION_CALL
2026-09-28 10:50:53 +08:00
Wesley Liddick 1df5c39311 Merge pull request #7562 from william-drakemond/fix/opencode-zen-deepseek-reasoning-placeholder
fix(openai): inject DeepSeek reasoning placeholder for OpenCode Zen upstream
2026-09-28 10:50:02 +08:00
Wesley Liddick b3c7038cf4 Merge pull request #7526 from HuntercodeT/fix/passthrough-model-discovery-merge
fix(gateway): let OpenAI passthrough accounts supplement model discovery instead of hiding mapped models
2026-09-28 10:49:01 +08:00
Wesley Liddick 8616d4e036 Merge pull request #7617 from korkin25/fix/responses-beta-upstream
fix(openai): preserve caller beta headers for HTTP Responses
2026-09-28 10:48:48 +08:00
Wesley Liddick 93c3da03eb Merge pull request #7555 from Yukinoshita-Yukino/fix/openai-auto-reset-credit-sched-projection
fix(scheduler): keep auto reset-credit fields in the scheduler projection
2026-09-28 10:27:00 +08:00
Achord Chan 68c2f5ae6d fix(anthropic): preserve requested structured outputs beta 2026-09-28 10:10:47 +08:00
Wesley Liddick ff6b85e30b Merge pull request #7615 from UnlastingR/fix/codex-ws-context-rollover-upstream
fix(openai): break websocket chain on context rollover
2026-09-28 10:05:05 +08:00
Wesley Liddick 79c18ec836 Merge pull request #7609 from feeeei/fix/client-disconnect-499
fix(gateway): 客户端断开统一标记 499,不再被记成 502/200 上游错误
2026-09-28 10:04:07 +08:00
Wesley Liddick 1ff2bd6d95 Merge pull request #7573 from Yukinoshita-Yukino/fix/channel-image-price-inherit-when-unset
fix(billing): keep catalog image prices when channel pricing leaves them unset
2026-09-28 09:57:02 +08:00
Wesley Liddick f592b336a3 Merge pull request #7595 from wucm667/feat/issue-7587-allowlist-glob
feat(groups): support wildcards anywhere in model allowlists
2026-09-28 09:55:55 +08:00
Wesley Liddick 37f35b12ea Merge pull request #7524 from Achordchan/fix/plaza-video-rate
fix(model-plaza): apply independent video billing multipliers
2026-09-28 09:55:34 +08:00
Wesley Liddick e4b34446af Merge pull request #7619 from SavitarC/fix/account-stats-long-context-billing
fix(billing): 账号统计成本按 OpenAI 账号开关计入长上下文价
2026-09-28 09:55:19 +08:00
Wesley Liddick 90ae81e9f8 Merge pull request #7624 from wucm667/fix/issue-7620-future-quota-reset
fix(openai): retain quota pause until known future reset
2026-09-28 09:54:52 +08:00
Wesley Liddick 9375e288ba Merge pull request #7571 from SPHINX998/fix/openai-probe-model-unavailable-inconclusive
fix(openai): keep Responses support unknown when the probe model is missing
2026-09-28 09:54:31 +08:00
Wesley Liddick 06f70a6232 Merge pull request #7572 from SPHINX998/fix/openai-alpha-search-require-completion
fix(openai): require a successful completion before billing an alpha search fallback
2026-09-28 09:54:12 +08:00
Wesley Liddick 6906062187 Merge pull request #7597 from far5man/codex/opus55-openrouter-alias-main
fix: recognize OpenRouter Claude Opus 5.5 alias for billing
2026-09-28 09:53:49 +08:00
ResidualBlood 6b40449c3f fix(antigravity): exclude Signature-only events from meaningful data check
MALFORMED_FUNCTION_CALL responses from upstream contain only
thoughtSignature without any text, thinking, or tool calls.
The previous fix still treated delta.Signature and block.Signature
as meaningful data, so the empty-stream failover was never triggered.

- Remove Signature from isMeaningfulAntigravityCompatEvent checks
- Let hasMeaningfulData() also consult processor.HasContent() as
  a secondary guard for edge cases
2026-09-27 13:15:56 +00:00
wucm667 02088be2d1 fix(openai): retain quota pause until known future reset 2026-09-27 14:56:38 +08:00
SavitarC fd7af1d462 fix(billing): honor long-context gate in account stats pricing 2026-09-27 12:12:25 +08:00
korkin25 8b95dc94f8 fix(openai): preserve Responses multi-agent beta 2026-09-27 09:57:51 +08:00
UnlastingR 7626110d0b fix(openai): break websocket chain on context rollover 2026-09-26 07:45:10 -04:00
Yujun KimandClaude Opus 5.5 3c5ea297ff fix(billing): keep zero-cost usage log when Free Fast has no pricing
RecordUsage treats a missing price as zero cost on the first pass, but
the Free Fast Standard re-evaluation returned the same pricing error.
Requests in a free_openai_fast group sent at priority/fast tier to a
model without pricing therefore lost their usage log entirely.

Ignore pricing-unavailable errors on the Standard pass and keep the
zero-cost entry, matching the first pass.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-09-26 20:11:03 +09:00
feeeei f4f8ff04db fix(gateway): 客户端断开统一标记 499,不再被记成 502/200 上游错误
客户端在响应提交前断开时,只有排队/换号阶段和 OpenAI 的兜底写入会标 499。
Claude 非流式与 Gemini 的 Chat Completions / Messages / Responses 兼容入口
在等待上游时断开,由 GatewayHandler.ensureForwardErrorResponse 补写 502;
Gemini 原生入口什么都不写,落成 200 并被 ops 记成一条 provider 的
"Recovered upstream error";Antigravity 回写 502。这些行都按上游失败计入
SLA 与上游错误数。

- GatewayHandler.ensureForwardErrorResponse 补客户端断开检查,与 OpenAI 版一致
- GatewayHandler 的 Messages / ChatCompletions / Responses 与 Gemini 原生入口:
  首次选号失败时先判断客户端断开;ChatCompletions / Responses 循环开头改用
  failoverClientGone,不再直接 return 落成 200
- Gemini 原生入口:forward 失败时补 499 标记;并发槽错误改走
  concurrencyErrorResponse 统一映射(断开 499、排队满/超时 429、其余 503)
- Antigravity 上游响应前断开回写 499;googleapi 与 Antigravity 的 Google 错误
  格式把 499 映射为 CANCELLED
- OpenAI alpha search 非 failover 错误分支补断开检查
- Claude / Gemini / OpenAI 传输层错误:请求 context 已取消时不记 ops 上游事件
- ops:IgnoreContextCanceled 开启时跳过没有上游错误上下文的 499;带上游错误的
  499(上游失败后客户端没等到换号结果就离开)仍按上游失败落库

计费不变:Gemini 流式中途断开的用量处理不在本次范围。新增 handler/service
单测覆盖各入口的断开场景与 ops 过滤规则。
2026-09-26 17:13:10 +08:00
Achord Chan 510b680fca Merge upstream main and retain upstream scheduler test fix 2026-09-26 15:47:52 +08:00
ResidualBlood 52c2f4cb10 fix(antigravity): trigger failover retry on empty stream caused by MALFORMED_FUNCTION_CALL 2026-09-26 05:29:18 +00:00
far5man 630639f747 test: preserve Opus 5.5 thinking for OpenRouter alias 2026-09-25 07:01:27 +02:00
far5man 11ffee5543 fix: recognize OpenRouter Claude Opus 5.5 alias for billing 2026-09-25 06:56:40 +02:00
Yuan LiuandClaude Opus 5.5 85810365a6 fix(auto-reset): stop re-querying confirmed no-credit accounts and back off after query failures
With the credit config now visible to the scheduler, every scheduling pass
notifies the auto-reset worker for paused credit-enabled accounts.
evaluateAccount forced an upstream query whenever the reset threshold was
reached, regardless of the last result. Accounts with no credits left and
7d >= 95% were therefore polled on /wham/usage every ~15-20s across
instances (previously once a minute by the scanner), and during upstream
outages the failed state caused the same storm.

- At the reset threshold, skip the forced query while a no_credit result is
  still within the state TTL (10 min); re-query once it expires so newly
  granted credits are still picked up.
- After a query-stage failure (RESET_CREDIT_QUERY_FAILED,
  USAGE_SNAPSHOT_WRITE_FAILED, RESET_CREDIT_DETAILS_UNAVAILABLE), wait one
  minute before querying again, including snapshot-staleness triggers.
- Redemption-stage failures are unaffected and still retry immediately with
  the same idempotency key; accounts with available credits still query and
  redeem immediately.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-09-25 12:17:11 +08:00
wucm667 f0ebc183e1 feat(groups): allow wildcard positions in model allowlist 2026-09-25 09:12:09 +08:00
SPHINX998andClaude Opus 5.5 1d126c833b fix(openai): require a successful completion before billing an alpha search fallback
The personal-access-token /v1/alpha/search fallback turned any upstream
Responses SSE body into a 200 search result and billed one web search call.
Error, response.failed and response.incomplete events, a response.completed
with a non-success status or without its response, a truncated stream and a
bare [DONE] were all accepted, returning partial text as a successful and
billable search.

Parse the stream strictly: those shapes now return an error, so nothing is
written to the client and no WebSearchCalls result is produced. A successful
response.completed keeps the existing output and citation handling. The
existing PAT metadata test now feeds a Responses SSE stream, which is what
that upstream path returns.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-09-24 15:54:15 +08:00
SPHINX998andClaude Opus 5.5 9461196864 fix(openai): keep Responses support unknown when the probe model is missing
The API-key Responses capability probe treated every 404 as conclusive and
stored responses_supported=false. The probe model is the first mapped
upstream model in lexical order, so on third-party accounts it is often an
auxiliary entry such as codex-auto-review; a model_not_found answer then
permanently routed the account away from /v1/responses although the
endpoint works.

A 400/404 whose error code/type or message says the probe model does not
exist or is unavailable is now inconclusive, so the stored capability is not
overwritten. Plain 404/405 still mean the endpoint is missing. The probe
also prefers a general gpt-* text model (not an image model) among the
mapped models before falling back to the first one.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-09-24 15:50:00 +08:00